Continuous infusion of a nonlethal dose of Escherichia coli lipopolysaccharide (LPS) (0.5 mg/kg) in#dced early (3 h) accumulation of polymorphonuclear leukocytes (PMNL) in rat liver followed by later (30 h) greater extravasation of mononuclear phagocytes (MNP) (E. B. Rodriguez de Turco and J. A. Spitzer, J. Leukocyte Biol. 48:488-494, 1990). Nonparenchymal liver cells from rats treated for 3 and 30 h with LPS were recovered by centrifugal elutriation, yielding a 23-ml/min fraction (endothelial cells) and a 45-ml/min fraction (PMNL, Kupffer cells, and MNP), and compared for their capacity for basal and agonist-stimulated superoxide (02 ) production. Stimulation with phorbol myristate acetate and opsonized zymosan caused a dose-dependent release Of 02 from the 45-ml/min fraction derived from rats treated for 3 h with saline, but not from the 23-ml/min fraction. Further purification of the 45-ml/min fraction by discontinuous density gradient centrifugation into a Kupifer and a PMNL fraction revealed that most of the agonist-induced 02 release was generated by infiltrating PMNL at this early time point of LPS infusion. By 30 h of LPS infusion, although enhancement of the phorbol-12-myristate-13-acetateand opsonized zymosan-stimulated release of 02 was observed in the 45-ml/min fraction, but not in the 23-ml/min fraction, the maximum release of 02was smaller than that observed in the rats treated for 3 h. Our results support the following conclusions: (i) after a 3-h LPS infusion, PMNL found in the liver in increased numbers are also highly primed for agonist-stimulated release of 02while Kupifer cell priming is of a lesser extent; (ii) after a 30-h infusion of LPS, infiltrating MNP found in the liver in increased numbers are primed for agonist-induced 02 release, while priming of PMNL has diminished; (iii) at both 3 and 30 h of LPS infusion, liver endothelial cells are not significantly primed for agonist-stimulated 02 release; and (iv) in vivo priming by LPS infusion at both 3 and 30 h was not reversed by the experimental method used for cell recovery (ca. 3 h), thus suggesting that in vivo LPS priming of 02 release may ultimately lead to severe impairment of liver function and metabolism observed during endotoxemia and sepsis if not therapeutically blocked at an early time point.